Inter-Jurisdiction Migration and the Fiscal Policies of Local Governments
(April 3, 2019)
Darong Dai, Dennis W. Jansen, and Liqun Liu
Abstract
This paper first analyzes a fiscal-policy game between two jurisdictions connected
by mutual migration and obtains two main results. (i) As the mutual migration
intensifies, both jurisdictions in the Nash equilibrium choose more public
consumption, less public investment and more total spending, and finance the total
spending entirely with debt. (ii) While the Nash equilibrium without any restriction
on local government debts is characterized by too much public consumption, too
little public investment and excessive debt, the first-best allocation can be achieved
through Nash play by imposing the restriction that public consumption should only
be financed by a contemporary tax and not by borrowing. The paper then goes on
to analyze a model with one-directional migration and obtains results on how
migration affects the fiscal policies of both the jurisdiction of migration destination
and the jurisdiction of migration origin.
Keywords: local government debt; migration; fiscal externalities; debt limits
JEL Classification: H23; H74
Darong Dai
Institute for Advanced Research
Shanghai University of Finance and Economics
Shanghai 200433
P.R. China
Dennis W. Jansen
Department of Economics and Private Enterprise Research Center
Texas A&M University
College Station, TX 77843
U.S.A.
Liqun Liu (corresponding author)
Private Enterprise Research Center
Texas A&M University
College Station, TX 77843
U.S.A.
(979)845-7723
1
1. Introduction
The relationship between migrations and the fiscal policies of local governments has
drawn extensive economic analyses.1 It is properly emphasized in these studies that the causal
effect in this relationship could run in both directions. On the one hand, the high tax/debt policy
in a jurisdiction may cause its residents to migrate to another jurisdiction, a mechanism referred
to as tax/debt-driven migration. On the other hand, the presence of migration, or even the threat
to migrate, helps shape the fiscal policies in jurisdictions that are connected through migration.
In this paper we analyze the effects of inter-jurisdiction migration on the fiscal policies of
two local governments in a two-period setting. Local fiscal policies are determined in the first
period, and individuals in each jurisdiction face a probability of migrating to another jurisdiction
between the two periods. The residents in a jurisdiction have a tendency to run excessive debt
because, once moving to the other jurisdiction, they are no longer responsible for the debt
repayment in their original jurisdiction. In this aspect, our analysis is closely related to those that
have previously examined the debt-increasing effect of migration, including Daly (1969), Oates
(1972), Bruce (1995) and Schultz and Sjostrom (2001, 2004), among others.2 However, our
analysis goes beyond those earlier analyses by making a distinction between government
spending on public consumption (which produces a publicly-provided good in period 1) and
government spending on public investment (which produces a publicly-provided good in period
2). With a positive probability of out-migrating in the future, residents in a jurisdiction view
public investment in their jurisdiction as less valuable than when migration is
1 See, for example, Mirrlees (1982), Wilson (1982), Wildasin (1994), Leite-Monteiro (1997), Lehmann et al. (2014)
and Dai (2017). 2 The focus of our analysis is therefore on the “horizontal fiscal externalities” through inter-jurisdiction migration,
compared to analyses of “vertical fiscal externalities” that have a focus on the fiscal-policy interactions between a
national government and a sub-national government (Besley and Rosen 1998, Dahlby and Wilson 2003, and
Aronsson 2010).
2
impossible. Therefore, spending on public consumption and spending on public investment
should be treated as separate variables in the model of the fiscal policies of local governments
that are connected by migration.
Making a distinction between these two types of government spending facilitates
obtaining two new, clear-cut theoretical findings within a two-jurisdiction, two-period model in
which initially identical residents in each jurisdiction choose both public consumption and public
investment in period 1, before knowing their migration status in period 2. First, as the between-
jurisdiction migration intensifies, both jurisdictions in the Nash equilibrium incur more spending
on public consumption, less spending on public investment, and more total spending. Further,
this total spending will equal the level of government debt, because in the Nash equilibrium both
jurisdictions opt for debt financing in the presence of migration. Compared to the earlier studies
on the effect of migration on local governments' debt levels that do not treat public consumption
and public investment separately, the result above explicitly attributes the increased debt level
due to migration to increased public consumption (or more precisely, government spending
incurred to provide goods in the current period). Indeed, in response to a heightened probability
of migration, public investment (or more precisely, government spending incurred in the present
to provide goods in the future period) actually decreases exactly when the total government
spending increases.3
Second, while the Nash equilibrium without any restriction on local government debt is
characterized by too much public consumption, too little public investment and excessive debt,
the first-best allocation can be achieved through Nash play by imposing the restriction that public
3 For example, in a political economy with local elections, Schultz and Sjostrom (2004) find that migration leads to
over-accumulation of local public debt. One difference is that they do not consider public goods in different periods,
and hence cannot analyze the different effects of migration placed on public consumption and public investment.
3
consumption should only be financed by a contemporary tax and not by borrowing.4 This result
provides a theoretical justification, based on migration, for the widespread practice of national
governments imposing debt limits on local governments.5 In contrast, a balanced-budget
restriction, which prohibits government borrowing, would not achieve the first-best allocation.
This result is the same in spirit as that obtained by Bassetto and Sargent (2006) in a model of
overlapping generations without migration. They find that whenever demographics do not imply
Ricardian equivalence, a "golden rule" that separates capital and ordinary account budgets and
finances only capital spending with debt is approximately optimal.
In the real world, and especially when the jurisdictions are countries rather than
subnational regions within a country, migration is often one-directional, where one of the two
linked jurisdictions is the migration destination (e.g., the U.S., Australia or Canada) and the other
is the migration origin (e.g., Puerto Rico, Mexico, China or India). So the paper goes on to
analyze how the intensity of one-directional migration affects the fiscal policies of the two
jurisdictions. It finds that migration’s effects on the migration destination are very similar to
those obtained in the model with mutual migration, and that migration’s effects on the migration
origin critically depend on how responsive the fiscal policies of the migration destination are to
migration.
The main results are established in Section 2 in a model with two identical jurisdictions,
separable utility functions, a perfect dichotomy between public consumption and public
investment, no cross-jurisdiction debt holding, and exogenous migration. The focus of this
section is on the effects of mutual migration on two linked jurisdictions’ fiscal policies. Section
4 While this restriction says nothing about how the spending on public investment should be financed, both
jurisdictions will opt for financing public investment exclusively with debt in the Nash equilibrium. 5 For discussions of debt limits imposed by various national/federal governments, see Bird and Slack (1983),
Aronson and Hilley (1986) and Mathews (1986).
4
3 extends the analysis to the asymmetric case where the two linked jurisdictions/countries can be
of different sizes and migration is one-directional. The analysis in this section addresses
migration’s effects on two linked jurisdictions/countries’ fiscal policies in situations where one
jurisdiction/country is the destination of migration and the other is the origin of migration.
Section 4 shows that the results obtained in Section 2 can be generalized to the case of non-
separable utility functions, the case where the dichotomy between public consumption and public
investment is imperfect (i.e. the case where the publicly-provided goods are durable), the case of
cross-jurisdiction debt holding, and the more general case of endogenous migration. Section 5
concludes.
2. A Fiscal-Policy Game between Two Identical Jurisdictions with Mutual Migration
We construct a fiscal-policy game between two identical jurisdictions, denoted A and B.
These jurisdictions have the same number of residents, among other things. Individuals are
identical in preferences as well as in earning abilities. Suppose that each individual lives for two
periods with income 1y in period 1 and 2y in period 2, and each individual has a lifetime utility
function
1 1 1 1 2 2 2 2( ) ( ) ( ) ( )u c g G u c g G , (1)
where 1 2 and c c are respectively the private consumption in periods 1 and 2, and 1 2 and G G are
respectively the publicly-provided goods in periods 1 and 2, where 1 2 and G G are per-capita
amounts. Note that the publicly-provided goods in this paper are “private” or “rival” goods,
rather than “public” or “non-rival” goods. While in general publicly-provided goods can be
either “private” or “public” goods, we focus on publicly-provided “private” goods because the
5
present paper is mainly concerned with government financing in the presence of migration, not
with the optimal provision of non-rival “public” goods.6
All four functions in (1) are assumed to be strictly increasing and strictly concave. The
lifetime utility function of (1) is both timely separable and separable in the private consumption
and the publicly-provided good in the same period. For the issues analyzed in the present paper,
this “double separability” assumption is not overly restrictive, but it greatly simplifies the
technical aspect of the analysis.7
We focus on the decisions of the representative individual in jurisdiction A, because
those of the representative individual in jurisdiction B are exactly symmetric. In period 1, the
representative individual in jurisdiction A decides on the amount invested in productive capital
(i.e., private saving) Ak ( 0Ak would indicate borrowing instead of holding capital), public
consumption 1
AG and public investment 2
AG , all the amounts being stated in per-capita terms.
Note that variables 1
AG and 2
AG serve a dual purpose. On the output side, 1 2 and A AG G are
publicly-provided goods in period 1 and period 2, respectively. One implicit assumption – that
1 2 and A AG G are non-durable goods – is made here for a sharp interpretation of the results.8 On
the input side, 1 2 and A AG G are both government spending in period 1, and the important
difference is that 1
AG is used to provide publicly available goods and services in period 1
whereas 2
AG is used to provide publicly available goods and services in period 2. Based on the
6 Nevertheless, the main results of the paper obtained in Sections 2 and 4 still hold when
1 2 and G G are publicly-
provided “public” goods (with appropriate interpretation of functions 1g and 2g ).
7 In Section 4, we consider a more general lifetime utility function 1 1 2 2( , ) ( , )u c G v c G , where the standard
timely separability is maintained, but the private consumption and the publicly-provided good in the same period are
allowed to be non-separable.
8 In Section 4, we consider a more general situation where 1
AG is durable with various degrees.
6
input side interpretation, 1
AG is referred to as spending on public consumption (or simply public
consumption), and 2
AG as spending on public investment (or simply public investment).
Examples of 1
AG include spending on transfer payments, health care and education,9 whereas
examples of 2
AG include spending on infrastructure, R&D and environment.
The government can finance its total spending in period 1, 1 2 + A AG G , by a mix of tax
collections or debt finance. Suppose that A is the portion of the total spending that is financed
by debt. To focus on the main forces at work for the results derived in this model, we initially
assume that the debt issued by a jurisdiction is exclusively held by the residents of that
jurisdiction.10 Then the representative individual holds bonds in the amount of 1 2+A A
A G G ,
which will be repaid to him with interest in period 2. Regardless of the specific mix of tax and
debt, however, the representative individual’s private consumption in period 1 is
1 1 1 2
A A A Ac y k G G ,
and this will be true with or without a positive probability of migration.
At the beginning of period 2, a fixed portion, denoted p, of the population in jurisdiction
A will be randomly selected to migrate to jurisdiction B. At the same time, the same fixed
portion of the population in jurisdiction B migrate to jurisdiction A, so that the number of
individuals in each jurisdiction will not change after the process of mutual migration. Therefore,
we initially assume that the probability of migration is exogenous – that is, people migrate
9 Even though spending on education is usually treated as investment, it is “public consumption” according to the
terminology used in this paper because the (capitalized) benefits of education accrue to the individuals who receive
the education. 10 In Section 4, we relax this assumption to allow some share of one jurisdiction’s debt to be held by residents of the
other jurisdiction. It is shown that the symmetric Nash equilibrium is characterized by the same set of conditions
regardless of how much cross-jurisdiction debt holding is allowed.
7
between jurisdictions for reasons that are unrelated to local fiscal policies – to focus on the
externalities, through migration, of one jurisdiction’s fiscal policy on another’s. Reasons for
people to change residence locations even in the absence of tax/debt related motivations may
include marriage, a better job opportunity, a hobby or schooling. Nevertheless, this “exogenous
migration” assumption is not critical, and we will show that the results obtained in this section
still hold in a more general model with endogenous probability of migration.11
If staying put in period 2 (with probability 1-p), the representative individual will pay for
the debt repayment in jurisdiction A, and also receive utility from the public investment in
jurisdiction A. In particular, his private consumption is
,
2 2 1 2 1 2 2(1 ) ( + )(1 ) ( + )(1 ) (1 )A S A A A A A A
A Ac y k r G G r G G r y k r ,
where r is the rate of return earned on productive capital as well as government bonds, and the
debt repayment the government owes him and the tax required to make the payment cancel out.
This is why in the absence of migration an individual is indifferent between tax or bond finance
of period 1 government spending.
If migrating from jurisdiction A to B in period 2 (with probability p), the representative
individual will be subject to the government in jurisdiction B, paying taxes in period 2 for the
debt incurred by jurisdiction B, and receiving benefit from public investment that was
undertaken in jurisdiction B. In particular, his private consumption is
,
2 2 1 2 1 2(1 ) ( + )(1 ) ( + )(1 )A M A A A B B
A Bc y k r G G r G G r ,
11 In Section 4, we allow the probability of migrating to another jurisdiction to be a function of both the difference in
the level of publicly-provided goods in period 2 and the difference in the debt level between the two jurisdictions.
We show that the comparative statics results on the symmetric Nash equilibrium are qualitatively the same as those
obtained in this section.
8
where 1 2 and B BG G are the per-capita public consumption and per-capita public investment in
jurisdiction B, respectively.
Note that regardless of what jurisdiction B chooses (including the value of B ), and
regardless of the values of Ak , 1
AG and 2
AG , the representative individual’s ,
2
A Mc is maximized
when 1A , whereas 1
Ac and ,
2
A Sc are unaffected by the choice of A . Therefore, in period 1
the representative individual in jurisdiction A will always choose to finance the entire total
spending in period 1 with debt, that is 1A is part of jurisdiction A’s optimal fiscal policy.
Intuitively, with a positive probability of migrating to jurisdiction B and not being responsible
for debt retirement in jurisdiction A, the representative individual in jurisdiction A views debt
financing as less costly than contemporary-tax financing. The same argument leads to 1B .
Therefore, the representative individual in jurisdiction A chooses 1 2, and A A Ak G G to
maximize his expected lifetime utility
1 1 1 2 1 1 2 2 2 2
2 2 1 2 1 2 2 2
( ) (1 ) (1 ) (1 ) ( )
(1 ) ( )(1 ) ( )(1 ) ( ),
A A A A A A
A A A B B B
u y k G G g G p u y k r p g G
pu y k r G G r G G r pg G
(2)
taking 1 2 and B BG G as given. The first-order conditions with respect to
1 2, and A A Ak G G are
respectively
1 1 1 2 2 2
2 2 1 2 1 2
1 1 1 2 1 1 2 2 1 2 1 2
1 1 1 2 2 2
(1 )(1 ) (1 )
(1 ) (1 ) ( )(1 ) ( )(1 ) 0
( ) (1 ) (1 ) ( )(1 ) ( )(1 ) 0
(1 ) (
A A A A
A A A B B
A A A A A A A B B
A A A
u y k G G r p u y k r
r pu y k r G G r G G r
u y k G G g G r pu y k r G G r G G r
u y k G G p g G
2 2 1 2 1 2) (1 ) (1 ) ( )(1 ) ( )(1 ) 0.A A A A B Br pu y k r G G r G G r
(3)
9
It is straightforward to show that the corresponding Hessian matrix is negative definite, and
therefore the second-order condition for the expected lifetime utility maximization problem is
satisfied.
2. 1. The Effect of Migration on the Symmetric Nash Equilibrium
Imposing the symmetry conditions, that * * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G ,
and rearranging the three equations in (3), the symmetric Nash equilibrium ( * * *
1 2, ,k G G ) is
determined by the following equations:
* * * *
1 1 1 2 2 2
* * * *
1 1 1 2 1 1
* * * *
1 1 1 2 2 2
(1 ) (1 ) 0
(1 ) ( ) 0
( ) 0.
u y k G G r u y k r
p u y k G G g G
u y k G G g G
(4)
Proposition 1. The comparative statics of the symmetric Nash equilibrium ( * * *
1 2, ,k G G ) with
respect to p are * * * **
1 2 1 2( )0, 0, 0 and 0
dG dG d G Gdk
dp dp dp dp
.
Proof: See the appendix.
Proposition 1 says that as between-jurisdiction migration (p) increases, residents in both
jurisdictions hold less productive capital (*
0dk
dp ), both jurisdictions provide more public
consumption (*
1 0dG
dp ) and less public investment (
*
2 0dG
dp ), and both jurisdictions incur more
debt to finance a larger total spending in period 1 (* *
1 2( )0
d G G
dp
). These results are quite
intuitive. With a larger chance to move to another jurisdiction later, the representative individual
in each jurisdiction views public investment in his own jurisdiction as less valuable and debt
financing less costly. Therefore, public investment in each jurisdiction becomes smaller
10
(*
2 0dG
dp ), and both jurisdictions incur more debt (
* *
1 2( )0
d G G
dp
). The combination of more
total government spending and less public investment implies an increase in public consumption
(*
1 0dG
dp ). As for people holding less productive capital (
*
0dk
dp ), it is the implication of
government bonds being a perfect substitute for productive capital and people holding more
government bonds.
The key to the results in Proposition 1 is that an increase in outbound migration causes
both the costs of debt-finance of current and prepaid future services and the benefits of future
services to decrease. Therefore, migration's effects on a jurisdiction's fiscal policy obtained here
are very similar to the effects of mortality or politicians' short-term bias on the fiscal policy of a
jurisdiction without migration.
2.2. The First-Best Allocation and Optimal Financing of Government Spending
The first-best allocation in both jurisdictions, denoted 1 2
ˆ ˆ ˆ( , , )k G G , maximizes
1 1 1 2 1 1 2 2 2 2( ) (1 ) ( )u y k G G g G u y k r g G . (5)
1 2ˆ ˆ ˆ( , , )k G G is determined by the following first-order conditions:
1 1 1 2 2 2
1 1 1 2 1 1
1 1 1 2 2 2
ˆ ˆˆ ˆ (1 ) (1 ) 0
ˆ ˆ ˆ ˆ( ) 0
ˆ ˆ ˆ ˆ( ) 0.
u y k G G r u y k r
u y k G G g G
u y k G G g G
(6)
We can make two immediate observations regarding the relationship between the first-
best allocation 1 2
ˆ ˆ ˆ( , , )k G G and the Nash equilibrium allocation * * *
1 2( , , )k G G in the last
subsection. First, it is clear by comparing (4) and (6) that the two allocations are identical when
11
p = 0. This result is easy to understand because, in our model, the only factor that causes the
Nash equilibrium allocation to be suboptimal is the externalities of one jurisdiction’s fiscal
policy on another via migration. Second, individuals in both jurisdictions are strictly better off
under the first-best allocation 1 2
ˆ ˆ ˆ( , , )k G G than under the Nash equilibrium allocation
* * *
1 2( , , )k G G when p > 0. To see this, note that the expected lifetime utility of (2) is
* * * * * *
1 1 1 2 1 1 2 2 2 2( ) (1 ) ( )u y k G G g G u y k r g G (7)
at the Nash equilibrium allocation * * *
1 2( , , )k G G , and (7) is simply the lifetime utility function (5)
evaluated at * * *
1 2( , , )k G G , whereas (5) is maximized at the first-best allocation 1 2
ˆ ˆ ˆ( , , )k G G by
definition.
As pointed out earlier in this section, the Nash equilibrium has the entire government
spending in period 1 (i.e., 1 2
A AG G per-capita in jurisdiction A) solely financed by debt issue, i.e.
by government borrowing. Absent any restrictions on how the two kinds of government
spending are financed (by tax or debt), both jurisdictions would opt for complete debt financing.
In the rest of this section we investigate the possibility of the first-best allocation being
implemented – via Nash equilibrium – by imposing some restriction on the financing of
government spending.
First, consider the restriction that no government borrowing is allowed in both
jurisdictions. The analysis below shows that this financing restriction does not implement the
first-best allocation when there is a positive probability of migration. In this case, the entire
period-1 government spending in jurisdiction A, 1 2
A AG G on the per-capita basis, is financed by
12
a tax in period 1. Then, the expected lifetime utility of the representative individual in
jurisdiction A becomes
1 1 1 2 1 1 2 2 2 2 2 2( ) (1 ) (1 ) ( ) ( )A A A A A A Bu y k G G g G u y k r p g G pg G . (8)
The difference between (8) and (2) lies in that the total government spending 1 2
A AG G is entirely
financed by a tax in period 1 in (8), and entirely financed by debt in (2).
Maximizing (8), the first-order conditions with respect to 1 2, and A A Ak G G are respectively
1 1 1 2 2 2
1 1 1 2 1 1
1 1 1 2 2 2
(1 ) (1 ) 0
( ) 0
(1 ) ( ) 0.
A A A A
A A A A
A A A A
u y k G G r u y k r
u y k G G g G
u y k G G p g G
(9)
Note that the best choices of 1 2, and A A Ak G G by jurisdiction A are independent of the choices by
jurisdiction B, implying that the resulting Nash equilibrium is also a dominant strategy
equilibrium.
It is obvious from comparing (9) with (4) and (6) that, when p = 0, the dominant strategy
equilibrium allocation under the no-debt financing is identical to both the Nash equilibrium
allocation under the all-debt financing and the first-best allocation, but when p > 0, the three
allocations diverge from one another. Therefore, the presence of migration (and the externalities
of one jurisdiction’s fiscal policy on another) is not only the source of the suboptimal fiscal
policies in the two jurisdictions, but also the source of the nonequivalence between debt and tax
financing.
Next, consider the financing restriction that public consumption must be entirely financed
by the contemporary tax. The analysis below shows that this financing restriction implements
13
the first-best allocation when there is a positive probability of migration. In this case, the
expected lifetime utility of the representative individual in jurisdiction A is
1 1 1 2 1 1 2 2 2 2
2 2 2 2 2 2
( ) (1 ) (1 ) (1 ) ( )
(1 ) (1 ) (1 ) ( ).
A A A A A A
A A B B
u y k G G g G p u y k r p g G
pu y k r G r G r pg G
(10)
The representative individual chooses 1 2, and A A Ak G G to maximize (10), taking
2
BG as
given. The first-order conditions with respect to 1 2, and A A Ak G G are respectively
1 1 1 2 2 2 2 2 2 2
1 1 1 2 1 1
1 1 1 2 2 2 2 2 2 2
(1 )(1 ) (1 ) (1 ) (1 ) (1 ) (1 ) 0
( ) 0
(1 ) ( ) (1 ) (1 ) (1 ) (1 ) 0.
A A A A A A B
A A A A
A A A A A A B
u y k G G r p u y k r r pu y k r G r G r
u y k G G g G
u y k G G p g G r pu y k r G r G r
Imposing the symmetry conditions, that * * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G , and
rearranging the three equations above, the symmetric Nash equilibrium ( * * *
1 2, ,k G G ) under the
restriction of no-debt financing of public consumption is determined by the following equations:
* * * *
1 1 1 2 2 2
* * * *
1 1 1 2 1 1
* * * *
1 1 1 2 2 2
(1 ) (1 ) 0
( ) 0
( ) 0.
u y k G G r u y k r
u y k G G g G
u y k G G g G
(11)
Noting that (11) and (6) are identical, we immediately have the following proposition.
Proposition 2. The first-best allocation is achieved in the Nash equilibrium by imposing the
restriction that public consumption must be financed by the contemporary tax.
It is important to note that the financing restriction that implements the first-best
allocation only requires all of public consumption to be financed by the current tax, and does not
further require that all of public investment be financed by debt. Nevertheless, requiring all of
public consumption to be financed by current tax alone is sufficient for efficiency exactly
because the equilibrium choice for the financing of public investment will exclusively be debt.
14
3. A Fiscal-Policy Game between Two Jurisdictions with One-Directional Migration
We modify the fiscal-policy game in the last section by allowing the two jurisdictions to
be of different sizes and by imposing an additional assumption that migration is one-directional.
A fiscal-policy game with these features can be used to analyze migration’s effects on fiscal
policies in situations where one of the two linked jurisdictions/countries is the destination of
migration (e.g., the U.S., Australia or Canada) and the other is the origin of migration (e.g.,
Puerto Rico, Mexico, China or India).12
The initial population size of jurisdiction A (the migration destination) is normalized to 1
while that of jurisdiction B (the migration origin) is denoted by parameter (0, )n . Assume
that individuals in jurisdiction A will always stay, whereas individuals in jurisdiction B migrate
to jurisdiction A, with probability p, in period 2.13 As in the last section, individuals are assumed
to be identical within and between jurisdictions
3.1 Optimal Fiscal Policy of Jurisdiction A (the Destination of Migration)
The first- and second-period consumptions of the representative individual born in
jurisdiction A are, respectively, given by
1 1 1 2
1 22 2 1 2
2 1 2
( + )(1 )(1 ) ( + )(1 )
1
(1 ) ( + )(1 ).1
A A A A
A AA A A A A
A
A A A
A
c y k G G
G G rc y k r G G r
np
npy k r G G r
np
(12)
12 As is demonstrated in this section, a game with these features turns out to be mathematically manageable, even
though some derivations can be quite involved. In contrast, a more general fiscal-policy game with non-identical
jurisdictions and two-way migration would be intractable because of the many interdependent endogenous variables. 13When n > 1, migration is from a country with a larger population to a country with a smaller population (e.g. from
China to Canada or India to Australia). When n < 1, on the other hand, migration is from a country with a smaller
population to a country with a larger population (e.g. from Puerto Rico or Mexico to the (Mainland) U.S.).
15
In period 1, the representative individual in jurisdiction A decides on the amount invested in
productive capital Ak , per-capita public consumption 1
AG , per-capita public investment 2
AG and
the fraction of period-1 public spending that is financed by borrowing A to maximize
21 1 1 1 2 2 2( ) ( ) ( )
1
AA A A G
u c g G u c gnp
. (13)
Note that migration from jurisdiction B has two opposing effects on Jurisdiction A.
First, it helps pay for the debt retirement in jurisdiction A, lowering the tax burden born by the
representative individual in jurisdiction A. This effect is reflected by the additional component
in period-2 consumption, 1 2( + )(1 )1
A A
A
npG G r
np
, which would be zero if p = 0. Second,
migration from jurisdiction B competes for period-2 public resources in jurisdiction A, as period-
2 per-capita consumption of publicly-provided goods is 2
1
AG
np, which is smaller than
2
AG due to
migration.
Because migration from jurisdiction B makes debt financing less costly than
contemporary-tax financing in jurisdiction A, 1A must be part of the optimal fiscal policy in
jurisdiction A. Therefore, jurisdiction A’s problem reduces to choosing Ak , 1
AG and 2
AG to
maximize (13), subject to (12) and 1A .
The first-order conditions with respect to 1 2, and A A Ak G G are respectively
1 1 1 2 2 2 1 2
1 1 1 2 1 1 2 2 1 2
21 1 1 2 2
(1 ) (1 ) ( + )(1 ) 01
( ) (1 ) (1 ) ( + )(1 ) 01 1
1(1
1 1
A A A A A A
A A A A A A A
AA A A
npu y k G G r u y k r G G r
np
np npu y k G G g G r u y k r G G r
np np
Gu y k G G g
np np
2 2 1 2) (1 ) ( + )(1 ) 0.
1 1
A A Anp npr u y k r G G r
np np
(14)
16
It is straightforward to show that the corresponding Hessian matrix is negative definite, and
therefore the second-order condition for the expected lifetime utility maximization problem is
satisfied.
Note that from (14), (np) – the number of migrants to jurisdiction A relative to the size of
its original population – can be treated as a single parameter in jurisdiction A’s optimization
problem. The proposition below states how jurisdiction A’s optimal fiscal policy responses to
changes in this parameter.
Proposition 3. The comparative statics of the optimal choices by jurisdiction A –
1 2, andA A Ak G G – with respect to the number of incoming migrants (np) are
10, 0,( ) ( )
AA dGdk
d np d np 1 2( )
and 0( )
A Ad G G
d np
.
Proof: See the appendix.
Proposition 3 says that as the number of incoming migrants increases, residents in
jurisdiction A hold less productive capital ( 0( )
Adk
d np ), spend more on public consumption (
1 0( )
AdG
d np ) and incur more debt to finance a larger total spending in period 1 ( 1 2( )
0( )
A Ad G G
d np
).
The explanation for these results is straightforward. With more incoming migrants to help with
debt retirement, jurisdiction A has an incentive to incur more debt ( 1 2( )0
( )
A Ad G G
d np
) to finance a
larger amount of public consumption in period 1 ( 1 0( )
AdG
d np ). Moreover, because people in
jurisdiction A hold more government bonds which are a perfect substitute for productive capital,
they hold less productive capital ( 0( )
Adk
d np ).
17
These results for the migration destination are very similar to those for the two identical
jurisdictions in a model of mutual migration stated in Proposition 1. This seems to suggest that
the in-migration alone provides much of the explanation for the comparative statics results
obtained within the model of mutual migration.
Nevertheless, one result in Proposition 1, namely *
2 0dG
dp , does not have a counterpart in
Proposition 3. In other words, there does not exist a clear-cut result for the sign of 2
( )
AdG
d np. The
reason for this is also readily understandable. Whereas more incoming migrants make the costs
of financing public investment 2
AG lower, they also further dilute the benefits from the public
investment. These two effects work in opposite directions, rendering the sign of 2
( )
AdG
d np
theoretically ambiguous.
3.2 Optimal Fiscal Policy of Jurisdiction B (the Origin of Migration)
The period-1 private consumption of the representative individual in jurisdiction B is
1 1 1 2
B B B Bc y k G G , (15)
where Bk is the individual’s investment in productive capital, 1
BG is per-capita public
consumption and 2
BG is per-capita public investment, all in jurisdiction B.
The period-2 private consumption of the representative individual in jurisdiction B
depends on whether the individual stays in jurisdiction B or migrates to jurisdiction A in period
2. If staying put in period 2 (with probability 1-p), the individual will pay for the debt repayment
in jurisdiction B, resulting in his private consumption being
18
, 1 22 2 1 2
2 1 2
( + )(1 )(1 ) ( + )(1 )
1
(1 ) ( + )(1 )1
B BB S B B B B
B
B B B
B
G G rc y k r G G r
p
py k r G G r
p
, (16)
where B is the fraction of period-1 public spending in jurisdiction B that is debt-financed.
If migrating to jurisdiction A in period 2 (with probability p), the representative
individual will be subject to the government in jurisdiction A, paying taxes in period 2 to retire
the debt incurred by jurisdiction A (noting that jurisdiction A optimally opts for 1A ).
Therefore, his private consumption in this case is
, 1 22 2 1 2
( + )(1 )(1 ) ( + )(1 )
1
A AB M B B B
B
G G rc y k r G G r
np
. (17)
Taking 1 2 and A AG G as given, jurisdiction B chooses
1, ,B B
B k G and 2
BG to maximize
, ,2 21 1 1 1 2 2 2 2 2 2( ) ( ) (1 ) ( ) ( )
1 1
B AB B B S B MG G
u c g G p u c g p u c gp np
, (18)
subject to (15) – (17). Note that, unlike the optimization problems in Section 2 and Subsection
3.1, the representative individual in jurisdiction B does not necessarily want 1B because
although there is a probability p for him to migrate to jurisdiction A, hence not being responsible
for the debt retirement in jurisdiction B, there is also a probability 1–p for him to stay put, hence
picking up the debt-retiring taxes that are left behind by those who migrated to jurisdiction A.
The first-order conditions with respect to 1, ,B B
B k G and 2
BG are respectively
, ,
1 2 2 2 1 2 2 2
, ,
1 1 2 2 2 2
, ,
1 1 1 1 2 2 2 2
, ,21 1 2 2 2 2 2
(1 )( + ) (1 )( + ) 0
(1 )(1 ) (1 ) 0
(1 ) (1 ) 0
(1 ) (1 )1
B B B S B B B M
B B S B M
B B B S B M
B B
BB B S B M
B B
p r G G u c p r G G u c
u c r p u c r pu c
u c g G r p u c r p u c
Gu c g r p u c r p u c
p
0.
(19)
19
It is straightforward to show that the corresponding Hessian matrix is negative definite, and
therefore the second-order condition for the expected lifetime utility maximization problem is
satisfied.
The first equation in (19) implies , ,
2 2
B S B Mc c . So (19) can be simplified to
1 2 1 2
1 1 1 2 2 2 1 2
1 1 1 2 1 1
21 1 1 2 2
1( + ) ( + ) 0
1
(1 ) (1 ) ( + )(1 ) 01
0
0,1
B B A A
B
B B B B B B
B
B B B B
BB B B
pG G G G
np
pu y k G G r u y k r G G r
p
u y k G G g G
Gu y k G G g
p
(20)
which is the set of equations that fully determine the optimal 1, ,B B
B k G and 2
BG , where 1 2+A AG G
and other parameters (n and p in particular) are taken as given.
Importantly, parameters p and n have impacts on the optimal 1, ,B B
B k G and 2
BG also
through 1 2+A AG G . Therefore, migration’s effects on the fiscal policies of jurisdiction B may
depend on how responsive jurisdiction A’s fiscal policies are to migration. The following
propositions about the comparative statics results for jurisdiction B indicate that this is indeed the
case. We first give the proposition about the comparative statics with respect to p.
Proposition 4. The comparative statics of jurisdiction B’s optimal choices with respect to p are
characterized as follows: If
1 2
1 2
+ 1
+ 1 1
A A
A A
d G G p p n
dp G G p np
, (21)
then 0Bd
dp
, 20, 0,
BB dGdk
dp dp and
1 2( )0
B B
Bd G G
dp
.
Proof: See the appendix.
20
The left-hand side of (21) is the elasticity of 1 2+A AG G with respect to p, and the right-hand
side of (21) is increasing in both p and n. So condition (21) simply requires the elasticity of
1 2+A AG G with respect to p to be large enough relatively to the level of migration (the larger the p
or n, the higher the level of migration). Condition (21) highlights the role played by jurisdiction
A’s response to migration in determining the optimal response to migration by jurisdiction B.
While condition (21) is only sufficient, not necessary, for the comparative statics results
in Proposition 4, it is both sufficient and necessary for the result of 1 2( )
0
B B
Bd G G
dp
. In
other words, condition (21) is identical to requiring that the debt level in jurisdiction B increase
as the probability of migration increases.
We now give the proposition about the comparative statics with respect to n.
Proposition 5. The comparative statics of jurisdiction B’s optimal choices with respect to n are
characterized as follows: 0Bd
dn
, 1 20, 0, 0,
B BB dG dGdk
dn dn dn and
1 2( )0
B B
Bd G G
dn
if
and only if
1 2
1 2
+
+ 1
A A
A A
d G G n np
dn G G np
. (22)
Proof: See the appendix.
The left-hand side of (22) is the elasticity of 1 2+A AG G with respect to n, and the right-hand
side of (22) is increasing in np. So condition (22) simply requires the elasticity of 1 2+A AG G with
respect to n to be large enough relatively to the level of migration (the larger the np, the higher
the level of migration). Note that condition (22) is both sufficient and necessary. So the optimal
21
responses of jurisdiction B to an increase in n critically depend on how responsive jurisdiction A
is to the increase in n.
4. Generalizations to the Non-Separable Utility Functions, Durable Goods, Cross-
Jurisdiction Debt Holding, and Endogenous Migration
In the last two sections, we assumed that the utility functions are separable in the private
consumption and the publicly-provided good in the same period, that there is a perfect
dichotomy between public consumption and public investment, that the debt issued by a
jurisdiction is exclusively held by the residents of that jurisdiction, and that the probability of
migration is exogenous to the fiscal policies of local governments, in order to simplify the
technical aspect of the analysis and to gain sharp insights. In this section, we generalize the
benchmark analysis in Section 2 of a fiscal-policy game between two identical jurisdictions by
allowing the utility functions to be non-separable, some government spending to simultaneously
provide public consumption and public investment (i.e. the case of durable publicly-provided
goods), some share of one jurisdiction’s debt to be held by the residents of the other jurisdiction,
or the probability of migration to be endogenous.
4.1. Non-Separable Utility Functions
We replace the lifetime utility function of (1) with
1 1 2 2( , ) ( , )u c G v c G . (1’)
Note that this lifetime utility function includes (1) as a special case in which 0cG cGu v .
Moreover, it also includes the often-used lifetime utility function, 1 1 2 2( , ) ( , )u c G u c G , as a
special case where 2 2 2 2( , ) ( , )v c G u c G .14 Functions u and v are assumed to be twice
14 is typically treated as a preference parameter with more patient individuals having a larger value of , but it
may also include a longevity component in which a lower mortality rate corresponds to a larger value of .
22
differentiable, strictly increasing in both variables (i.e. , , , 0c G c Gu u v v ) and strictly concave
(i.e. , , , 0cc GG cc GGu u v v , 2( ) 0cc GG cGu u u and 2( ) 0cc GG cGv v v ). In addition, we assume
that c and G are weakly complements in both u and v (i.e. , 0cG cGu v ).15
Taking 1 2, and B B
BG G as given, the representative individual in jurisdiction A chooses
1 2, , and A A A
Ak G G to maximize
, ,
1 1 2 2 2 2( , ) (1 ) ( , ) ( , )A A A S A A M Bu c G p v c G pv c G , (2’)
where
1 1 1 2
,
2 2 1 2 1 2 2
,
2 2 1 2 1 2
(1 ) ( + )(1 ) ( + )(1 ) (1 )
(1 ) ( + )(1 ) ( + )(1 ).
A A A A
A S A A A A A A
A A
A M A A A B B
A B
c y k G G
c y k r G G r G G r y k r
c y k r G G r G G r
Obviously, the individual’s optimal choices should include 1A . Therefore, the above
lifetime utility maximization problem can be reduced to a maximization problem with respect to
1 2, andA A Ak G G , taking 1 2 andB BG G as given and letting 1A B . The first-order conditions
with respect to 1 2, and A A Ak G G are respectively
1 1 2 1 2 2
2 1 2 1 2 2
1 1 2 1 1 1 2 1
2 1 2 1 2 2
, (1 )(1 ) (1 ),
(1 ) (1 ) ( + )(1 ) ( + )(1 ), 0
, ,
(1 ) (1 ) ( + )(1 ) ( + )(1 ), 0
A A A A A A
c c
A A A B B B
c
A A A A A A A A
c G
A A A B B B
c
c
u y k G G G r p v y k r G
r pv y k r G G r G G r G
u y k G G G u y k G G G
r pv y k r G G r G G r G
u
1 1 2 1 2 2
2 1 2 1 2 2
, (1 ) (1 ),
(1 ) (1 ) ( + )(1 ) ( + )(1 ), 0.
A A A A A A
G
A A A B B B
c
y k G G G p v y k r G
r pv y k r G G r G G r G
(3’)
15 The assumption of complementarity between c and G is somewhat restrictive because it does not allow the two
goods to be substitutes. On the other hand, this assumption is satisfied by most often-used utility functions and
greatly simplifies the technical aspect of the analysis in this section.
23
It is straightforward to show that the corresponding Hessian matrix is negative definite, and
therefore the second-order condition for the expected lifetime utility maximization problem is
satisfied.
Imposing the symmetry conditions, that * * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G ,
and rearranging the three equations in (3’), the symmetric Nash equilibrium (* * *
1 2, ,k G G ) is
determined by the following equations:
* * * * * *
1 1 2 1 2 2
* * * * * * * *
1 1 2 1 1 1 2 1
* * * * * *
1 1 2 1 2 2
, (1 ) (1 ), 0
(1 ) , , 0
, (1 ), 0.
c c
c G
c G
u y k G G G r v y k r G
p u y k G G G u y k G G G
u y k G G G v y k r G
(4’)
Proposition 1’. Under the stated assumptions on non-separable utility functions u and v, the
comparative statics of the symmetric Nash equilibrium (* * *
1 2, ,k G G ) with respect to p are still
characterized by * **
1 20, 0 and 0dG dGdk
dp dp dp . In addition,
* *
1 2( )0
d G G
dp
if
(1 ) (1 ) (1 )cc cG GG cG cc cGu r v v r u r v v , which is particularly true when 0cGu .
Proof: See the appendix.
Now we demonstrate that Proposition 2 in Section 2 can also be generalized to the case of
non-separable utility functions. In the more general case, the first-best allocation in both
jurisdictions, denoted 1 2
ˆ ˆ ˆ( , , )k G G , maximizes
1 1 2 1 2 2, (1 ),u y k G G G v y k r G . (5’)
1 2ˆ ˆ ˆ( , , )k G G is thus determined by the following first-order conditions:
24
1 1 2 1 2 2
1 1 2 1 1 1 2 1
1 1 2 1 2 2
ˆ ˆˆ ˆ ˆ ˆ, (1 ) (1 ), 0
ˆ ˆˆ ˆ ˆ ˆ ˆ ˆ, , 0
ˆ ˆˆ ˆ ˆ ˆ, (1 ), 0.
c c
c G
c G
u y k G G G r v y k r G
u y k G G G u y k G G G
u y k G G G v y k r G
(6’)
Consider the effect on the symmetric Nash equilibrium from the financing restriction that
public consumption must be entirely financed by the contemporary tax. In this case, the expected
lifetime utility of the representative individual in jurisdiction A is
1 1 2 1 2 2
2 2 2 2
, (1 ) (1 ),
(1 ) (1 ) (1 ), .
A A A A A A
A A B B
u y k G G G p v y k r G
pv y k r G r G r G
(10’)
The representative individual chooses 1 2, and A A Ak G G to maximize (10’), taking 2
BG as given.
The first-order conditions with respect to 1 2, and A A Ak G G are respectively
1 1 2 1 2 2
2 2 2 2
1 1 2 1 1 1 2 1
1 1 2 1 2 2
2 2
, (1 )(1 ) (1 ),
(1 ) (1 ) (1 ) (1 ), 0
, , 0
, (1 ) (1 ),
(1 ) (1 ) (1
A A A A A A
c c
A A B B
c
A A A A A A A A
c G
A A A A A A
c G
A A
c
u y k G G G r p v y k r G
r pv y k r G r G r G
u y k G G G u y k G G G
u y k G G G p v y k r G
r pv y k r G
2 2) (1 ), 0.B Br G r G
Imposing the symmetry conditions, that * * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G , and
rearranging the three equations above, the symmetric Nash equilibrium (* * *
1 2, ,k G G ) under the
restriction of no-debt financing of public consumption is determined by the following equations:
* * * * * *
1 1 2 1 2 2
* * * * * * * *
1 1 2 1 1 1 2 1
* * * * * *
1 1 2 1 2 2
, (1 ) (1 ), 0
, , 0
, (1 ), 0.
c c
c G
c G
u y k G G G r v y k r G
u y k G G G u y k G G G
u y k G G G v y k r G
(11’)
Noting that (11’) and (6’) are identical, we immediately have the following proposition.
25
Proposition 2’. For the more general case with non-separable utility functions u and v, the first-
best allocation is achieved in the Nash equilibrium by imposing the restriction that public
consumption must be financed by the contemporary tax.
4.2. Durable Publicly-Provided Goods
A main feature of this paper is the distinction between spending on public consumption
and spending on public investment. The critical role played by this distinction is apparent in
Propositions 1 and 2 of Section 2. Proposition 1 says that, among other things, as mutual
migration intensifies, equilibrium public consumption increases and equilibrium public
investment decreases; Proposition 2 says that under the restriction that public consumption is
exclusively financed by the contemporary tax, the first-best allocation is implemented through
the Nash equilibrium.
A question then arises as to whether these two propositions can be generalized to the case
where some government spending simultaneously provides public consumption and public
investment, i.e., the case of durable publicly-provided goods.
In this case, the representative individual in jurisdiction A chooses 1 2, and A A Ak G G – and
for the same previously discussed reason the total spending 1 2
A AG G is financed by debt – to
maximize his expected lifetime utility
1 1 1 2 1 1 2 2 2 1 2
2 2 1 2 1 2 2 1 2
( ) (1 ) (1 ) (1 ) ( )
(1 ) ( )(1 ) ( )(1 ) ( ),
A A A A A A A
A A A B B B B
u y k G G g G p u y k r p g G G
pu y k r G G r G G r pg G G
(2”)
26
taking 1 2 and B BG G as given. In (2”), government spending 1G produces both public
consumption and public investment, and [0,1) measures the durability of publicly-provided
good 1G .16 Obviously, (2) in Section 2 is a special case of (2”) here where 0 .
From the same derivations as those in Section 2, the symmetric Nash equilibrium
(* * *
1 2, ,k G G ) is found to satisfy:
* * * *
1 1 1 2 2 2
* * * * * *
1 1 1 2 1 1 2 1 2
* * * * *
1 1 1 2 2 1 2
(1 ) (1 ) 0
(1 ) ( ) (1 ) ( ) 0
( ) 0.
u y k G G r u y k r
p u y k G G g G p g G G
u y k G G g G G
(4”)
Parallel to Proposition 1, we obtain the following result, the proof of which is omitted
because it is similar to that of Proposition 1.
Proposition 1”. The comparative statics of the symmetric Nash equilibrium (* * *
1 2, ,k G G ) with
respect to p are * * * * **
1 1 2 1 2( ) ( )0, 0, 0 and 0
dG d G G d G Gdk
dp dp dp dp
.
Note that * *
1 2( )0
d G G
dp
implies
*
2 0 dG
dp , given
*
1 0dG
dp . The result
* *
1 2( )0
d G G
dp
is easy to understand once it is recognized that
* *
1 2G G is the total amount of
publicly-provided good in period 2.
We next establish the parallel result to Proposition 2, for the case of durable 1G . The
first-best allocation, denoted 1 2
ˆ ˆ ˆ( , , )k G G , maximizes
1 1 1 2 1 1 2 2 2 1 2( ) (1 ) ( )u y k G G g G u y k r g G G . (5”)
16Note that
2G is strictly dominated by 1G if 1 . So we only consider [0,1) to give a role to 2G .
27
The first-order conditions are:
1 1 1 2 2 2
1 1 1 2 1 1 2 1 2
1 1 1 2 2 1 2
ˆ ˆˆ ˆ (1 ) (1 ) 0
ˆ ˆ ˆ ˆ ˆ ˆ( ) ( ) 0
ˆ ˆ ˆ ˆ ˆ( ) 0.
u y k G G r u y k r
u y k G G g G g G G
u y k G G g G G
(6”)
In order to implement the first-best allocation through the Nash play, consider the
financing restriction in both jurisdictions that a portion of government spending on 1G ,
1(1 )G , must be financed by the contemporary tax. The analysis below shows that this
financing restriction implements the first-best allocation. In this case, the expected lifetime
utility of the representative individual in jurisdiction A is
1 1 1 2 1 1 2 2 2 1 2
2 2 1 2 1 2 2 1 2
( ) (1 ) (1 ) (1 ) ( )
(1 ) ( )(1 ) ( )(1 ) ( ).
A A A A A A A
A A A B B B B
u y k G G g G p u y k r p g G G
pu y k r G G r G G r pg G G
(10”)
It can be shown that the symmetric Nash equilibrium, denoted (* * *
1 2, ,k G G ), is determined by the
following conditions:
* * * *
1 1 1 2 2 2
* * * *
1 1 1 2 1 1
* * * * *
1 1 1 2 2 1 2
(1 ) (1 ) 0
(1 ) ( ) 0
( ) 0.
u y k G G r u y k r
u y k G G g G
u y k G G g G G
(11”)
Noting that (11”) and (6”) are equivalent, we immediately have the following
proposition.
Proposition 2”. The first-best allocation is achieved in the Nash equilibrium by imposing the
restriction that a portion of government spending on 1G , 1(1 )G , must be financed by the
contemporary tax.
28
One implication from comparing Proposition 2 and Proposition 2” is that the public
financing institution that is able to implement the first-best allocation under non-durable
publicly-provided goods is different from that under durable goods. The durability of publicly-
provided good 1G makes spending on 1G simultaneously produce public consumption (which is
1G ) and public investment (which is 1G ). An efficient public financing institution only
requires that a portion of spending on 1G , 1(1 )G to be specific, be financed by the current
tax.
4.3. Cross-Jurisdiction Debt Holding
Again suppose that jurisdiction A can finance its total spending in period 1, 1 2 + A AG G , by
a mix of tax collections or debt finance, and that A is the portion of the total spending that is
financed by debt. However, instead of assuming that the per-capita debt amount 1 2+A A
A G G is
entirely held by the representative individual in jurisdiction A, we now assume that only some
share (0 1) of this debt is held by the representative individual. Then the representative
individual in jurisdiction A holds bonds issued by jurisdiction A in the amount of 1 2+A A
A G G
and bonds issued by jurisdiction B in the amount of 1 2(1 ) +B B
B G G . The representative
individual’s private consumption in period 1 is therefore
1 1 1 2 1 2 1 2(1 )( ) ( ) (1 ) ( )A A A A A A B B
A A Bc y k G G G G G G .
If staying put in period 2 (with probability 1-p), the representative individual will pay for
the debt repayment in jurisdiction A, and also receive utility from the public investment in
jurisdiction A. In particular, his private consumption is
,
2 2 1 2 1 2 1 2(1 ) ( )(1 ) (1 ) ( )(1 ) ( + )(1 )A S A A A B B A A
A B Ac y k r G G r G G r G G r ,
29
where r is the rate of return earned on productive capital as well as government bonds.
If migrating from jurisdiction A to B in period 2 (with probability p), the representative
individual will be subject to the government in jurisdiction B, paying taxes in period 2 for the
debt incurred by jurisdiction B, and receiving benefit from public investment that was
undertaken in jurisdiction B. In particular, his private consumption is
,
2 2 1 2 1 2 1 2(1 ) ( )(1 ) (1 ) ( )(1 ) ( + )(1 )A M A A A B B B B
A B Bc y k r G G r G G r G G r .
Note that regardless of what jurisdiction B chooses (including the value of B ), and
regardless of the values of Ak , 1
AG and 2
AG , the representative individual’s ,
2
A Mc is maximized
when 1A , whereas the constraint for the net present value of 1
Ac and ,
2
A Sc , or ,
21
1
A SA c
cr
,
remains 21 1 2( )
1
A A yy G G
r
, which is independent of A . Therefore, in period 1 the
representative individual in jurisdiction A will always choose to finance the entire total spending
in period 1 with debt, that is 1A . The same argument leads to 1B .
Therefore, the representative individual in jurisdiction A chooses 1 2, and A A Ak G G to
maximize his expected lifetime utility
1 1 1 2 1 2 1 1
2 2 1 2 1 2 2 2
2 2 1 2 1 2 2 2
( ) (1 )( ) ( )
(1 ) (1 ) (1 )( )(1 ) (1 )( )(1 ) (1 ) ( )
(1 ) ( )(1 ) ( )(1 ) ( ),
A A A B B A
A A A B B A
A A A B B B
u y k G G G G g G
p u y k r G G r G G r p g G
pu y k r G G r G G r pg G
(2”’)
taking 1 2 and B BG G as given. The first-order conditions with respect to 1 2, and A A Ak G G are
respectively
30
1 1 1 2 1 2 2 2 1 2 1 2
2 2 1 2 1 2
1 1 1 2 1 2 1 1
( ) (1 )( ) (1 )(1 ) (1 ) (1 )( )(1 ) (1 )( )(1 )
(1 ) (1 ) ( )(1 ) ( )(1 ) 0
( ) (1 )( ) ( )
(
A A A B B A A A B B
A A A B B
A A A B B A
u y k G G G G r p u y k r G G r G G r
r pu y k r G G r G G r
u y k G G G G g G
2 2 1 2 1 2
2 2 1 2 1 2
1 1 1 2 1 2 2 2 1 2
1 )(1 )(1 ) (1 ) (1 )( )(1 ) (1 )( )(1 )
(1 ) (1 ) ( )(1 ) ( )(1 ) 0
( ) (1 )( ) (1 )(1 )(1 ) (1 ) (1 )(
A A A B B
A A A B B
A A A B B A A
r p u y k r G G r G G r
r p u y k r G G r G G r
u y k G G G G r p u y k r G G
1 2
2 2 2 2 1 2 1 2
)(1 ) (1 )( )(1 )
(1 ) ( ) (1 ) (1 ) ( )(1 ) ( )(1 ) 0.
A B B
A A A A B B
r G G r
p g G r p u y k r G G r G G r
(3”’)
Imposing the symmetry conditions, that * * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G ,
and rearranging the three equations in (3”’), the symmetric Nash equilibrium (* * *
1 2, ,k G G ) is
determined by the following equations:
* * * *
1 1 1 2 2 2
* * * *
1 1 1 2 1 1
* * * *
1 1 1 2 2 2
(1 ) (1 ) 0
(1 ) ( ) 0
( ) 0.
u y k G G r u y k r
p u y k G G g G
u y k G G g G
(4”’)
Note that (4”’) is identical to (4). That is, the Nash equilibrium is characterized by the same set
of conditions regardless of the value of , the share of the debt issued by a jurisdiction that is
held by the residents of the same jurisdiction.
4.4. Endogenous Migration
Here we allow the probability of migration to depend on levels of debt and of public
goods in our original model with two identical jurisdictions that we discussed in Section 2. We
show that the symmetric Nash equilibrium under endogenous probability of migration is
characterized by a similar set of conditions to those in equation (4), so that all the results
obtained in Section 2 continue to hold in this more general case.
For reasons already discussed in Section 2, the representative individual in each
jurisdiction will always choose to finance the entire total spending in period 1 with debt.
31
Therefore, the representative individual in jurisdiction A makes decisions in period 1 on savings
Ak , spending on public consumption 1
AG , and spending on public investment 2
AG . By saving
Ak and lending 1 2
A AG G to the government, the representative individual’s private and public
consumption in period 1 are respectively 1 1 1 2
A A A Ac y k G G and 1
AG .
At the beginning of period 2, a portion of the population in jurisdiction A will be
randomly selected to migrate to jurisdiction B. The probability of out-migrating faced by the
representative individual in jurisdiction A is determined by debt and by publicly provided goods
according to
2 2 2 2 1 2 1 2, , ( ) ( )A B A A B B A A A B Bp p G G D D p G G G G G G ,
where AD and BD are respectively the debt level (in per-capita terms) in period 1 in jurisdiction
A and B, and the probability function ,p satisfies 0, 0 0p , 1 0p , 2 0p ,
11 0p and
22 0p . (Symmetrically, the probability of out-migrating faced by the representative individual
in jurisdiction B is given by 2 2 2 2 1 2 1 2, , ( ) ( )B A B B A A B B B A Ap p G G D D p G G G G G G .)
The probability function ,p captures the idea that the motivation for moving to the other
jurisdiction is strengthened by the level of period-2 publicly-provided goods in the other
jurisdiction relative to that in the native jurisdiction, as well as by the debt level in the native
jurisdiction relative to that in the other jurisdiction. Note also that 0, 0 0p captures a
baseline level of migration that is active even if the fiscal policies in the two jurisdictions are
identical. Thus the results here encompass those reported above in Section 2 as a special case.
If staying put in period 2 (with probability 1Ap ), the representative individual will pay
for the debt repayment in jurisdiction A, and also receive utility from the public investment in
32
jurisdiction A. Specifically, his private consumption and his share of publicly-provided goods in
period 2 are respectively
, 1 22 2 1 2 2 1 2
( + )(1 ) 1(1 ) ( + )(1 ) (1 ) ( + )(1 ) 1
1 1
A AA S A A A A A A
B A B A
G G rc y k r G G r y k r G G r
p p p p
and 2
1
A
B A
G
p p .
If migrating from jurisdiction A to B in period 2 (with probability Ap ), the representative
individual will be subject to the government in jurisdiction B, paying taxes in period 2 for the
debt incurred by jurisdiction B, and receiving benefit from public investment that was
undertaken in jurisdiction B. Specifically, his private consumption and his share of publicly-
provided goods in period 2 are respectively
, 1 22 2 1 2
( + )(1 )(1 ) ( + )(1 )
1
B BA M A A A
A B
G G rc y k r G G r
p p
and 2
1
B
A B
G
p p .
Therefore, the representative individual in jurisdiction A chooses 1 2, and A A Ak G G to
maximize his expected lifetime utility
1 1 1 2 1 1 2 2 1 2
2 1 2 22 2 2 1 2 2
1( ) (1 ) (1 ) ( + )(1 ) 1
1
( + )(1 )(1 ) (1 ) ( )(1 ) ,
1 1 1
A A A A A A A A
B A
A B B BA A A A A A
B A A B A B
u y k G G g G p u y k r G G rp p
G G G r Gp g p u y k r G G r p g
p p p p p p
taking 1 2 and B BG G as given. Note that 2 2 1 2 1 2, ( ) ( )A B A A A B Bp p G G G G G G and
2 2 1 2 1 2, ( ) ( )B A B B B A Ap p G G G G G G in the above objective function.
33
The first-order conditions with respect to 1 2, and A A Ak G G are given in the appendix. Then
imposing the symmetry requirements on the first order conditions -- that
* * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G – leads to
* * * *
1 1 1 2 2 2
* * * * * *
1 1 1 2 1 1 2 2 2 2
* * *
2 1 2 2 2
* * * *
1 1 1 2 2 2 2 1 2
(1 ) (1 ) 0,
(1 ) ( ) 2(1 2 ) (0,0) ( )
2(1 2 ) (0,0)(1 ) (1 ) 0,
(1 ) (1 ) ( ) 2(1 2 ) (0,0) (0,0)
u y k G G r u y k r
p u y k G G g G p p G g G
p p r G G u y k r
p u y k G G p g G p p p G
* *
2 2
* * *
2 1 1 2 2 2
( )
2(1 2 ) (0,0) (0,0) (1 ) (1 ) 0,
g G
p p p r G G u y k r
where (0, 0)p p . To simplify the analysis, we assume 2 1(0,0) (0,0)p p , which would hold,
for example, if the probability functions take the more restrictive form of
2 2 2 2, ( ) ( )A B A A B B B A Ap p G G D D f G D G D and
2 2 2 2, ( ) ( )B A B B A A A B Bp p G G D D f G D G D , where f satisfies 0f , 0f .
These probability functions imply that migrants are attracted by the net savings ( 2G D ) of the
other jurisdiction relative to that of one’s native jurisdiction. Then the characterizing conditions
for the symmetric Nash equilibrium simplify to
* * * *
1 1 1 2 2 2
* * * * * *
1 1 1 2 1 1 2 1 2 2
* * * *
1 1 1 2 2 2
(1 ) (1 ) 0,
(1 ) ( ) 2(1 2 ) (1 ) (1 ) 0,
( ) 0.
u y k G G r u y k r
p u y k G G g G p p r G u y k r
u y k G G g G
where 2 2(0,0)p p . Comparative statics results for
* * *
1 2, andk G G with respect to p are shown
in the appendix to have the signs * * * **
1 2 1 2( )0, 0, 0 and 0
dG dG d G Gdk
dp dp dp dp
. Therefore the
results in Section 2 hold in the more general case of endogenous migration probabilities.
34
5. Concluding Summary
We construct a two-jurisdiction, two-period model to study the effects of the between-
jurisdiction migration on the fiscal policies of local governments. In period 1, each jurisdiction
decides how much to spend on public consumption and how much on public investment. Each
jurisdiction can finance this total spending with a mix of debt and contemporary taxation. At the
beginning of period 2, a portion of the residents in each jurisdiction are randomly selected to
migrate to another jurisdiction, and, as a result, both receive benefits from public investment in
the new jurisdiction and also incur responsibility for the debt repayment in the new
jurisdiction.
Two main findings are obtained in this context. First, as the level of migration increases,
both the spending on public consumption and total spending (which is 100% financed by
government borrowing in the Nash equilibrium) increase, whereas spending on public
investment decreases. The implication is that increased human mobility is another reason for
excessive government debt. Second, the first-best allocation can be obtained as a Nash
equilibrium outcome by imposing the restriction that public consumption should be exclusively
financed with contemporary taxation. This provides a theoretical justification for various
national governments imposing debt limits on local governments.
The paper also considers a model with one-directional migration and obtains results on
how migration affects the fiscal policies of both the jurisdiction of migration destination and the
jurisdiction of migration origin. It finds that migration’s effects on the migration destination are
very similar to those obtained in the model with mutual migration, and that migration’s effects
on the migration origin critically depend on how responsive the fiscal policies of the migration
destination are to migration.
35
References
Aronson, R., Hilley, J., 1986. Financing State and Local Governments. The Brookings
Institution: Washington, DC.
Aronsson, T., 2010. Optimal income taxation and decentralized fiscal federalism. Regional
Science and Urban Economics 40 (4), 187-195.
Bassetto M., Sargent, T.J., 2006. Politics and efficiency of separating capital and ordinary
government budgets. Quarterly Journal of Economics 121(4), 1167-1210.
Besley, T.J., Rosen, H.S., 1998. Vertical externalities in tax setting: evidence from gasoline and
cigarettes. Journal of Public Economics 70, 383-398.
Bird, R., Slack, N.E., 1983. Urban Public Finance in Canada. Butterworths: Toronto.
Bruce, N., 1995. A fiscal federalism analysis of debt policies by sovereign regional
governments. Canadian Journal of Economics 28, 195-206.
Dahlby, B., Wilson, L.S., 2003. Vertical fiscal externalities in a federation. Journal of Public
Economics 87, 917-930.
Dai, D., 2017. Voting over selfishly optimal income tax schedules with tax-driven migrations.
Working paper, November 2017.
Daly, G., 1969. The burden of the debt and future generations in local finance. Southern
Economic Journal 36, 44-51.
Lehmann, E., Simula, L., Trannoy, A., 2014. Tax me if you can! Optimal nonlinear income tax
between competing governments. Quarterly Journal of Economics 129(4), 1995-2030.
Leite-Monteiro, M., 1997. Redistributive policy with labour mobility across countries. Journal
of Public Economics 65 (2), 229-244.
Mirrlees, J.A., 1982. Migration and optimal income taxes. Journal of Public Economics 18(3),
319-341.
Oates, W.E., 1972. Fiscal Federalism. Harcourt, Brace, Jovanovich: New York.
Mathews, R., 1986. The Australian Loan Council: Co-ordination of public debt policies in a
federation. In Public Finance and Public Debt, B. Herber, ed. Wayne State University
Press: Detroit.
Schultz, C., Sjostrom, T., 2001. Local public goods, debt and migration. Journal of Public
Economics 80, 313-337.
Schultz, C., Sjostrom, T., 2004. Public debt, migration, and shortsighted politicians. Journal of
Public Economic Theory 6, 655-674.
Wildasin, D.E., 1994. Income redistribution and migration. Canadian Journal of Economics 27,
637-656.
Wilson, J.D. 1982. Optimal linear income taxation in the presence of emigration. Journal of
Public Economics 18, 363-379.
36
Appendix
Proof of Proposition 1
Differentiating the three equations in (4) with respect to p, respectively, taking * * *
1 2, ,k G G
as functions of p, yields
* **2 1 2
1 2 1 1
* **
1 21 1 1 1 1
* **
1 21 1 1 2
(1 ) 0
(1 ) (1 ) (1 )
( ) 0
dG dGdku r u u u
dp dp dp
dG dGdkp u p u g p u u
dp dp dp
dG dGdku u u g
dp dp dp
.
We have
2
1 2 1 1
1 1 1 1
1 1 1 2
2 2
1 1 2 2 2 2 1 1
(1 )
(1 ) (1 ) (1 )
(1 ) (1 ) (1 ) 0.
u r u u u
p u p u g p u
u u u g
u g r u g u g r p u g
Therefore,
1 1*
11 1 1 1 1 2
1 1 2
2
1 2 1*2 21 1
1 1 1 1 2 1 2 2 2
1 1 2
*
2
01
(1 ) (1 ) 0,
0
(1 ) 01
(1 ) (1 ) (1 ) (1 ) 0,
0
1
u uudk
u p u g p u u gdp
u u g
u r u udG u
p u u p u r u u u g r u gdp
u u g
dG
dp
2
1 2 1
211 1 1 1 1 2
1 1
* *21 2 1
1 2 2 2
(1 ) 0
(1 ) (1 ) (1 ) 0,
0
( )(1 ) 0.
u r u uu
p u p u g u r u u
u u
d G G uu g r u g
dp
Proof of Proposition 3
Rearranging the equations in (14), we have
37
1 1 1 2 2 2 1 2
1 1 1 2 1 1
21 1 1 2 2
(1 ) (1 ) ( + )(1 ) 01
1( ) 0
1
0.1
A A A A A A
A A A A
AA A A
npu y k G G r u y k r G G r
np
u y k G G g Gnp
Gu y k G G g
np
Differentiating the above equations with respect to (np), respectively, yields
2 2 2
2 1 21 2 1 2 1 2 1 2 22
1 1 1 1 21
(1 ) (1 ) (1 )(1 ) ( )
( ) 1 ( ) 1 ( ) (1 )
1 ( ) 1 ( ) 1
A AAA A
A AA
dG dGdk np r np r ru r u u u u u G G u
d np np d np np d np np
u u dG u dGdkg
np d np np d np np d
12
1 2 2 21 1 1 22
1
( ) (1 )
.( ) ( ) 1 ( ) (1 )
A A AA
unp np
dG g dG Gdku u u g
d np d np np d np np
We have
2 2
2
1 2 1 2 1 2
1 1 11
21 1 1
2
2 2
1 2 1 2 1 2 1 2
(1 ) (1 )(1 )
1 1
1 1 1
1
1 1(1 ) (1 )
1 1
A
np r np ru r u u u u u
np np
u u ug
np np np
gu u u
np
g g u r u r u u g gnp np
0.
Therefore,
38
2 2 2
1 2 2 1 2 1 22
1 11 12
2 22 1 12
22
1 2 2 1 22
(1 ) (1 ) (1 )( )
(1 ) 1 1
1 1
( ) (1 ) 1 1
(1 ) 1
(1 ) 1( )
(1 ) 11
A A
A
A
A
A A
A
r np r np rG G u u u u u
np np np
u udku g
d np np np np
G gg u u
np np
rG G u u g u
np np
1 21 1
2
2 12 1 1 22
2 22
1 2 1 2 2 1 22
1 112
10,
(1 )
(1 ) 1 1
(1 ) (1 )(1 ) ( )
(1 ) 1
1 1
( ) 1 (1 )
A
A A
A
A
g gg
np
g u np rG g u u
np np np
r np ru r u G G u u u
np np
dG uu
d np np np
1
2 21 2 12
2 221 2 1 1 2 2 1 1 2
1 23 4 3
2 22
1 2 1 2
2
1
(1 ) 1
(1 ) (1 )1(1 ) 0,
(1 ) (1 ) (1 )
(1 ) (1 )(1 )
1 (1
1
( )
A
A
A
A
A
u
np
G gu g u
np np
u g r G u u g r u u uu r u
np np np
np r ru r u u u
np np
dG
d np
1 2 22
1 11 12
21 1 22
2 2 222 1 2 2 1 2 2 1 1 2 1 2 1 2 1
1 22 4 3 2
( ))
1
1 1 (1 )
(1 )
(1 ) (1 ) (1 ) ( )1(1 )
(1 ) (1 ) (1 ) (1 )
A A
A
A A A A
A
G G u
u ug u
np np np
Gu u g
np
G g g r G u u g r u u u r G G u u gu r u
np np np np
22 21 2 1 1 2 2 2 1 2
1 2 1 23 4 2
1 2
2 2
2 1 2 2 1 2 1 2 1
4 2
,
(1 )(1 ) (1 )
(1 ) (1 ) (1 )( ) 1
( ) (1 ) (1 ) ( )
(1 ) (1 )
A A
A A
A A A A
u g r G u u g G g gu r u u r u
np np npd G G
d np r G u u g r G G u u g
np np
0.
Proof of Proposition 4
Differentiating the four equations in (20) with respect to p, respectively, yields
39
1 21 21 2 1 22
2 221 2 2 2 1 2
1 2 1
2
1 2
+1 1( + ) ( + )
1 (1 )
(1 ) ( + ) (1 )(1 )
1 1
(1 ) ( )
A AB BB B A AB
B
B B B BB
B B
B B
B
d G Gd dG dG p nG G G G
dp dp dp np dp np
r p G G u d r p u dG dGdku r u u
p dp dp p dp dp
r G G
2
2
1 21 1 1 1
1 2 21 1 1 2 22
(1 )
0
1.
1 (1 )
B BB
B B BB
u
p
dG dGdku u g u
dp dp dp
dG dG Gdku u u g g
dp dp p dp p
We have
1 2
2 2 221 2 2 2 2
1 2 1 1
1 1 1 1
1 1 1 2
2 22 2
1 2 1 1
1 2
0
(1 ) ( + ) (1 ) (1 )(1 )
1 1 1
0
10
1
(1 ) (1 )(1 )
1
B B
B B
B B
B B
B
B B
B B
G G
r p G G u r p u r p uu r u u u
p p p
u u g u
u u u gp
r p u r pu r u u u
p
G G
2
1 1 1 1
1 1 1 2
2
1 2 21 1 1 1
1 1 1 2
2 2
1 2 1 2 1 2 1 2 1
1
1
1
0(1 ) ( + )
11
1
1 1(1 ) (1 )
1 1
B BB B
B B
u
p
u u g u
u u u gp
r p G G uu u g u
p
u u u gp
G G u r u g g r u u gp p
2 0.g
(i) BB
B
d
dp
, where
40
1 2
1 22
2 2 221 2 2 2 2
1 2 1 12
1 1 1 1
22 1 1 1 22
+1 1( + ) 0
1 (1 )
(1 ) ( ) (1 ) (1 )(1 )
(1 ) 1 1
0
1
(1 ) 1
1
1
B
A A
A A
B B
B B
B B B
B
d G Gp nG G
np dp np
r G G u r p u r p uu r u u u
p p p
u u g u
Gg u u u g
p p
p
2 22 2 2
1 2 1 1
1 2
1 2 1 1 1 12
1 1 1 2
2 2
1 2 1 21 22
(1 ) (1 )(1 )
1 1+ 1
( + )(1 )
1
1
(1 ) ( + ) 1
(1 ) 1
B B
A A
A A
B B
B
r p u r p uu r u u u
p pd G G n
G G u u g unp dp np
u u u gp
r G G u ug g
p p
22
1 2 1 22(1 ) ,
1
B
BGu r u g g
p
where
2 22 2 2
1 2 1 1
1 1 1 1
1 1 1 2
(1 ) (1 )(1 )
1 1
0
1
1
B Br p u r p uu r u u u
p p
u u g u
u u u gp
can be readily checked.
So a sufficient condition for 0Bd
dp
is (21).
(ii) B
B
k
B
dk
dp
, where
41
1 2
1 2 1 22
2 2 2 2
1 2 2 1 2 2 2 21 12
1 1 1
22 1 12
+1 1( + )
1 (1 )
(1 ) ( + ) (1 ) ( ) (1 ) (1 )
1 (1 ) 1 1
0 0
10
(1 ) 1
B
A A
B B A A
B B
B B B B
B B B
k
B
d G Gp nG G G G
np dp np
r p G G u r G G u r p u r p uu u
p p p p
u g u
Gg u u
p p
2
2 2 2
1 2 2 2 21 12
1 2 1 1 1
22 1 1 22
1 2
1 22
2
1 2 2
(1 ) ( ) (1 ) (1 )
(1 ) 1 1
0
1
(1 ) 1
+1 1( + )
1 (1 )(1 ) ( + )
1
B B
B B B
B B
B
A A
A A
B B
g
r G G u r p u r p uu u
p p p
G G u g u
Gg u u g
p p
d G Gp nG G
np dp npr p G G u
p
1 1 1
22 1 1 22
2 2 2
1 2 2 2 21 12
1 2 1 1 1
22 1 1 22
2
1 2
0
1
(1 ) 1
(1 ) ( ) (1 ) (1 )
(1 ) 1 1
0
1
(1 ) 1
(1 ) ( + )
B B
B
B B
B B B
B B
B
B B
u g u
Gg u u g
p p
r G G u r p u r p uu u
p p p
G G u g u
Gg u u g
p p
r p G G
1 22 21 2 1 2 1 1 1 2 1 22 2
+1 1 1 1( + )
1 1 (1 ) 1 1 (1 )
A A BA A B
d G Gu Gp nG G u g u g g g g g
p np dp np p p p
where
2 2 2
1 2 2 2 21 12
1 1 1
22 1 1 22
(1 ) ( ) (1 ) (1 )
(1 ) 1 1
0 0
1
(1 ) 1
B B
B B B
B
r G G u r p u r p uu u
p p p
u g u
Gg u u g
p p
can be readily
checked.
So a sufficient condition for 0Bdk
dp is (21).
42
(iii) 22B
BG
B
dG
dp
, where
2
1 2
1 2 1 22
2 2 221 2 2 2 1 2 2
1 2 1 2
1 1 1
21 1 22
1
1 2
+1 10 ( + )
1 (1 )
(1 ) ( + ) (1 ) (1 ) ( )(1 )
1 1 (1 )
0 0
0(1 )
(1
B
A A
B B A A
B
B B B B
B B
G
B
B B
d G Gp nG G G G
np dp np
r p G G u r p u r G G uu r u u
p p p
u u g
Gu u g
p
u r
G G
2 22 2 1 2 2
2 1 2
1 1 1
21 1 22
21 21 2 2 2
1 2 1 1 1 22 2
(1 ) (1 ) ( ))
1 (1 )
0
(1 )
+(1 ) ( + ) 1 1( + )
1 (1 ) 1 (1 )
B B
B B
B
A AB B BA AB
r p u r G G uu u
p p
u u g
Gu u g
p
d G Gr p G G u G p nu g u g G G
p p np dp np
,
where
2 22 2 1 2 2
1 2 1 2
1 1 1
21 1 22
(1 ) (1 ) ( )(1 )
1 (1 )
0 0
(1 )
B B
B B
B
r p u r G G uu r u u
p p
u u g
Gu u g
p
can be readily checked.
So a sufficient condition for 2 0BdG
dp is (21)
(iv) From the first equation in (20), it immediately follows that 1 2( )
0
B B
Bd G G
dp
if and only if (21) holds.
Proof of Proposition 5
Differentiating the four equations in (20) with respect to n, respectively, yields
43
1 21 21 2 1 22
2 221 2 2 2 1 2
1 2 1
1 1
+1 (1 )( + ) ( + )
1 (1 )
(1 ) ( + ) (1 )(1 ) 0
1 1
A AB BB B A AB
B
B B B BB
B B
B
d G Gd dG dG p p pG G G G
dn dn dn np dn np
r p G G u d r p u dG dGdku r u u
p dn dn p dn dn
dku u
dn
1 21 1
1 21 1 1 2
0
10.
1
B B
B BB
dG dGg u
dn dn
dG dGdku u u g
dn dn p dn
We have
1 2
2 2 221 2 2 2 2
1 2 1 1
1 1 1 1
1 1 1 2
0
(1 ) ( + ) (1 ) (1 )(1 )
1 1 10,
0
10
1
B B
B B
B B
B B
B
G G
r p G G u r p u r p uu r u u u
p p p
u u g u
u u u gp
from the proof of Proposition 4.
We thus have
1 2
1 22
2 22 2 2
1 2 1 1
1 1 1 1
1 1 1 2
1 2
+1 (1 )( + ) 0
1 (1 )
(1 ) (1 )1 0 (1 )
1 1
0
10
1
+1 1 (1 )
1 (1
A A
A A
B B
B BB
B
A A
B
d G Gp p pG G
np dn np
r p u r p ud u r u u u
p pdn
u u g u
u u u gp
d G Gp p p
np dn
2 22 2 2
1 2 1 1
1 2 1 1 1 12
1 1 1 2
(1 ) (1 )(1 )
1 1
( + ))
1
1
B B
A A
r p u r p uu r u u u
p p
G G u u g unp
u u u gp
44
where
2 22 2 2
1 2 1 1
1 1 1 1
1 1 1 2
(1 ) (1 )(1 )
1 1
0
1
1
B Br p u r p uu r u u u
p p
u u g u
u u u gp
from the proof of Proposition 4.
So 0Bd
dn
if and only if (22) holds. It can be similarly shown that (22) is also the
sufficient and necessary condition for 1 20, 0, 0,B BB dG dGdk
dn dn dn and
1 2( )0
B B
Bd G G
dn
.
Proof of Proposition 1’
Differentiating the three equations in (4’) with respect to p, respectively, taking
* * *
1 2, ,k G G as functions of p, yields
* **2 1 2
* **
1 21
* **
1 2
(1 ) ( ) (1 ) 0
(1 ) (1 )( ) (1 )
(1 ) ( ) ( ) 0,
cc cc cc cG cG cc
cc cG cc cG cG GG cc cG
cc cG cc cG cc GG
dG dGdkr v u u u r v u
dp dp dp
dG dGdkp u u p u u u u p u u u
dp dp dp
dG dGdku r v u u u v
dp dp dp
.
in which we have used the fact that cG Gcu u and cG Gcv v based on Young’s Theorem. Let
2
2
2 2
(1 ) (1 )
(1 ) (1 )( ) (1 )
(1 )
( ) (1 ) (1 ) (1 )
( ) (1 ) (1 ) ( 2)
cc cc cc cG cG cc
cc cG cc cG cG GG cc cG
cc cG cc cG cc GG
cc GG cG cc cG GG cG
cc GG cG cc cG GG
r v u u u r v u
p u u p u u u u p u u
u r v u u u v
u u u r r v v v r v
v v v r p u p u u
0.
Therefore,
45
*
2
*
1
0 (1 )1
(1 )( ) (1 )
0
( ) (1 ) 0,
(1 ) 0 (1 )1
(1 ) (1 )
(1 ) 0
(1 )
cc cG cG cc
c cc cG cG GG cc cG
cc cG cc GG
ccc cG GG cG
cc cc cG cc
cc cG c cc cG
cc cG cc GG
c
u u r v udk
u p u u u u p u udp
u u u v
uu u v r v
r v u r v udG
p u u u p u udp
u r v u v
ur
2 2 2
2
*
2
2(1 ) (1 ) ( ) 0,
(1 ) 01
(1 ) (1 )( )
(1 ) 0
( )(1 ) (1 ) 0.
cc GG cG cc cc GG cG
cc cc cc cG
cc cG cc cG cG GG c
cc cG cc cG
ccc cG cc cG
v v r v u r v v v
r v u u udG
p u u p u u u u udp
u r v u u
uu u r r v v
Moreover,
* *
2 21 2( )(1 ) (1 ) (1 ) (1 ) ( ) 0c
cc cG GG cG cG cc cG cc GG
ud G Gu r v v r u v r v r v v v
dp
if (1 ) (1 ) (1 )cc cG GG cG cc cGu r v v r u r v v .
The FOCs for the Problem of the Representative Individual in Jurisdiction A, the Symmetric
Nash Equilibrium, and the Comparative Statics Results (for the Case of Endogenous Migration
in Section 4.4)
As is established in Section 4.4, the representative individual in jurisdiction A chooses
1 2, and A A Ak G G to maximize his expected lifetime utility
1 1 1 2 1 1 2 2 1 2
2 1 2 22 2 2 1 2 2
1( ) (1 ) (1 ) ( + )(1 ) 1
1
( + )(1 )(1 ) (1 ) ( )(1 ) ,
1 1 1
A A A A A A A A
B A
A B B BA A A A A A
B A A B A B
u y k G G g G p u y k r G G rp p
G G G r Gp g p u y k r G G r p g
p p p p p p
46
taking 1 2 and B BG G as given. Note that 2 2 1 2 1 2, ( ) ( )A B A A A B Bp p G G G G G G and
2 2 1 2 1 2, ( ) ( )B A B B B A Ap p G G G G G G in the above objective function.
The first-order conditions with respect to 1 2, and A A Ak G G are respectively the following:
1 1 1 2 2 2 1 2
1 22 2 1 2
1(1 )(1 ) (1 ) ( + )(1 ) 1
1
( + )(1 )(1 ) (1 ) ( )(1 ) 0,
1
A A A A A A A
B A
B BA A A A
A B
u y k G G r p u y k r G G rp p
G G rr p u y k r G G r
p p
1 1 1 2 1 1 2 2 2 1 2 1 2 2 2 1 2
2 2 1 2
1 2
1( ) , ( ) ( ) (1 ) ( + )(1 ) 1
1
1(1 ) (1 ) ( + )(1 ) 1
1
1(1 ) 1 ( + )(1 )
1
A A A A B A A A B B A A A
B A
A A A A
B A
A A
B A
u y k G G g G p G G G G G G u y k r G G rp p
p u y k r G G rp p
r G G rp p
2 2 2 1 2 1 2 2 2 2 1 2 1 2
2
22 2 2 1 2 1 2 2
2 2 2 2 1 2 1 2 2 22
2
, ( ) ( ) , ( ) ( )
(1 )
, ( ) ( )1
, ( ) ( )(1 )
1
A B B B A A B A A A B B
B A
AB A A A B B
B A
A A B B B A A BA
A
B A
p G G G G G G p G G G G G G
p p
Gp G G G G G G g
p p
G p G G G G G G p G GGp g
p p
2 1 2 1 2
2
1 22 2 2 1 2 1 2 2 2 1 2
1 22 2 1 2
1 2
, ( ) ( )
(1 )
( + )(1 ), ( ) ( ) (1 ) ( )(1 )
1
( + )(1 )(1 ) ( )(1 )
1
( + )(1 )(1 )
(1
A A A B B
B A
B BB A A A B B A A A
A B
B BA A A A
A B
B B
G G G G
p p
G G rp G G G G G G u y k r G G r
p p
G G rp u y k r G G r
p p
G G rr
2 2 2 1 2 1 2 2 2 2 1 2 1 22
22 2 2 1 2 1 2 2
2 22 2 2 2 1 2 1 2 2 2 2 12
, ( ) ( ) , ( ) ( ))
, ( ) ( )1
, ( ) ( ) , (1 (1 )
B A A A B B A B B B A A
A B
BB A A A B B
A B
B BA B A A A B B A B B
A B A B
p G G G G G G p G G G G G Gp p
Gp G G G G G G g
p p
G Gp g p G G G G G G p G G G
p p p p
2 1 2) ( ) 0,B A AG G G
and
47
1 1 1 2
2 2 2 1 2 1 2 1 2 2 1 2 1 2 2 2 1 2
2 2 1 2
1, ( ) ( ) , ( ) ( ) (1 ) ( + )(1 ) 1
1
1(1 ) (1 ) ( + )(1 ) 1
1
1(1 ) 1
1
A A A
B A A A B B B A A A B B A A A
B A
A A A A
B A
B A
u y k G G
p G G G G G G p G G G G G G u y k r G G rp p
p u y k r G G rp p
rp p
1 2
1 2 2 1 2 1 2 2 2 2 1 2 1 2 2 2 2 1 2 1 2 1 2 2 1 2 1 2
2
2 2 2 1 2 1
( + )(1 )
, ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( )
(1 )
, ( ) (
A A
A B B B A A A B B B A A B A A A B B B A A A B B
B A
B A A A B
G G r
p G G G G G G p G G G G G G p G G G G G G p G G G G G G
p p
p G G G G G
22 1 2 2 1 2 1 2 2
22
2 1 2 2 1 2 1 2 2 2 2 1 2 1 2 2 2 2 1 2 1 2 1 2 2 1 2
) , ( ) ( )1
(1 )1
1 , ( ) ( ) , ( ) ( ) , ( ) ( ) , ( )
AB B A A A B B
B A
AA
B A
B A A A B B B A A A B B B A A B A A A B B B A A A
GG p G G G G G G g
p p
Gp g
p p
p p G p G G G G G G p G G G G G G p G G G G G G p G G G G
1 2
2
1 22 2 2 1 2 1 2 1 2 2 1 2 1 2 2 2 1 2
1 22 2 1 2
( )
(1 )
( + )(1 ), ( ) ( ) , ( ) ( ) (1 ) ( )(1 )
1
( + )(1 )(1 ) ( )(1 )
1
(1 )
B B
B A
B BB A A A B B B A A A B B A A A
A B
B BA A A A
A B
G G
p p
G G rp G G G G G G p G G G G G G u y k r G G r
p p
G G rp u y k r G G r
p p
r
1 2
2
2 2 2 1 2 1 2 1 2 2 1 2 1 2 1 2 2 1 2 1 2 2 2 2 1 2 1 2
2 2 2 1 2 1 2 1 2 2 1
( + )(1 )
(1 )
, ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( )
, ( ) ( ) , (
B B
A B
B A A A B B B A A A B B A B B B A A A B B B A A
B A A A B B B A
G G r
p p
p G G G G G G p G G G G G G p G G G G G G p G G G G G G
p G G G G G G p G G G
22 1 2 2
2 22 2
2 2 2 1 2 1 2 1 2 2 1 2 1 2 1 2 2 1 2 1 2 2 2 2 1 2 1 2
) ( )1
1 (1 )
, ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) 0.
BA A B B
A B
B BA
A B A B
B A A A B B B A A A B B A B B B A A A B B B A A
GG G G g
p p
G Gp g
p p p p
p G G G G G G p G G G G G G p G G G G G G p G G G G G G
48
Then imposing the symmetry requirements, that * * *
1 1 1 2 2 2, and A B A B A Bk k k G G G G G G ,
on these first-order conditions leads to the following conditions that characterize the symmetric
Nash equilibrium,
* * * *
1 1 1 2 2 2
* * * * * *
1 1 1 2 1 1 2 2 2 2
* * *
2 1 2 2 2
* * * *
1 1 1 2 2 2 2 1 2
(1 ) (1 ) 0,
(1 ) ( ) 2(1 2 ) (0,0) ( )
2(1 2 ) (0,0)(1 ) (1 ) 0,
(1 ) (1 ) ( ) 2(1 2 ) (0,0) (0,0)
u y k G G r u y k r
p u y k G G g G p p G g G
p p r G G u y k r
p u y k G G p g G p p p G
* *
2 2
* * *
2 1 1 2 2 2
( )
2(1 2 ) (0,0) (0,0) (1 ) (1 ) 0,
g G
p p p r G G u y k r
where (0, 0)p p . Assuming 2 1(0,0) (0,0)p p , which is explained in Section 4.4, these Nash
equilibrium conditions reduce to
* * * *
1 1 1 2 2 2
* * * * * *
1 1 1 2 1 1 2 1 2 2
* * * *
1 1 1 2 2 2
(1 ) (1 ) 0,
(1 ) ( ) 2(1 2 ) (1 ) (1 ) 0,
( ) 0.
u y k G G r u y k r
p u y k G G g G p p r G u y k r
u y k G G g G
where 2 2(0,0)p p .
Differentiating the three equilibrium equations with respect to p, respectively, taking * * *
1 2, ,k G G as functions of p, yields
* **2 1 2
1 2 1 1
* **2 * 1 2
1 2 1 2 1 1 2 2 1
*
1 2 1 2
**
11 1 1 2
(1 ) 0,
(1 ) 2(1 2 ) (1 ) (1 ) 2(1 2 ) (1 ) (1 )
4 (1 ) ,
( )
dG dGdku r u u u
dp dp dp
dG dGdkp u p p r G u p u g p p r u p u
dp dp dp
u p r G u
dG dGdku u u g
dp dp
*
2 0.dp
.
Let
49
2
1 2 1 1
2 *
1 2 1 2 1 1 2 2 1
1 1 1 2
2 2
1 1 2 2 2 2 1 1
2 2 1 2
(1 )
(1 ) 2(1 2 ) (1 ) (1 ) 2(1 2 ) (1 ) (1 )
(1 ) (1 ) (1 )
2(1 2 ) (1 ) (
u r u u u
p u p p r G u p u g p p r u p u
u u u g
u g r u g u g r p u g
p p r u u g
2 2 2 *
1 2 2 2 2 1 2 1 21 ) (1 ) 2(1 2 ) (1 ) 0.r u u r u g p p r G u u g
Then,
1 1**
1 2 1 2 1 1 2 2 1
1 1 2
*
1 2 1 2 1
2
1 2 1*2 *1
1 2 1 2 1 2
01
4 (1 ) (1 ) 2(1 2 ) (1 ) (1 )
0
4 (1 )0,
(1 ) 01
(1 ) 2(1 2 ) (1 ) 4 (1 )
u udk
u p r G u p u g p p r u p udp
u u g
u p r G u u g
u r u udG
p u p p r G u u p r Gdp
*
1 2 1
1 1 2
*2 21 2 1 2
1 2 1 2 2 2
2
1 2 1*2 *2
1 2 1 2 1 1 2 2 1 2
(1 )
0
4 (1 )(1 ) (1 ) 0,
(1 ) 01
(1 ) 2(1 2 ) (1 ) (1 ) 2(1 2 ) (1 ) 4 (1 )
u p u
u u g
u p r G ur u u u g r u g
u r u udG
p u p p r G u p u g p p r u u p rdp
*
1 2
1 1
* 2
1 2 1 2 1 2
* * *21 2 1 2 1 2
1 2 2 2
0
4 (1 ) (1 )0,
( ) 4 (1 )(1 ) 0.
G u
u u
u p r G u r u u
d G G u p r G uu g r u g
dp